grobotsfandomcom-20200213-history
Strategies overview
War and peace There was one tournament where sides did not use weapons and just tried to accumulate as much biomass as possible. But apart from that all sucsesfull sides are extremely offencive and attack everything they see. Why is that? mostly because if you dont attack first, opponent will kill some of your cells. You cant predict what the opponent will do. When opponent attacks its already too late to react. and there is no way to store the data between games, so gour side cant learn which side is agressive and which is not. Even if there would be such way or if you would code it manually you cant know what the side name is. you can only guess by checking their weight and type. you cant know their max energy or max armor, and current values are not that helpful. In real world altruism appear only in cases where organism has enough memory to remember those who helped him and those who hurt him, and more memory this organism has, the higher chances of altruistic behaivor it spicie has. Looting eating corpses of enemies can be a good addition to mana, but the problem is that its practically impossible to predict how much mana you will get from looting. Unlike patrolling territory for mana, which gives stable income. Another problem is that you cant easily increase number of eaters and constructors when you hame few more corpses. The price for new gatherer may be more than the cost of the corps. But looting may be very useful. In example if you will collect corpses of your own fighters then your fighter may be twice as bad, but you still will be equal to your opponent. And if you will collect both corpses from your fighters and from opponents fighters then your fighters may be three times as bad, and you will be equal with opponent. so far no sides are collecting corpses on purpose. Economy Main concern may be in using a force fields or engine to collect mana. Because forcefields and engine cost nearly the same but force field allow you to load your eater to 100% in most cases forcefield is better. Engine may allow something like 50% load of your eater, because some time will be spent moving from one source to another. Another question is minimum and maximum eater per cell. Because mana looses 0.1 energy per turn your eater has to be at least 0.4 per mana source under your cell to collect 80% of the available mana. and to load your eater fully you have to store several mana sources under your cell to avoid waiting. this leads to minimum eater per unit of approximately 1.2-2 (3-5 mana sources). any smaler value will alow mana to evaporate for no good reason. because some energy is spent on weapons, engines, forcefields, sensors, etc, your eater has to be 20-50% bigger than your constructor. Next question is energy storage VS constructing unit. if you have excessive energy storage your cell will not start to use autoconstructor when energy is not nearly full. this may lead to lag at the start. So if you are using large energy storage it makes sense to manually start bulding units. multicellular VS situational Or in other words does it makes sense to try to make organs or focus on trying to make cell behave correctly on its own? So far all succesful sides was using situational behaivor, where each cell doesnt care much about other cells and simply does it main task. Reason is that optimal cell size of about 30 force users to make only 1-3 cells at the seed stage. its not enough to gain any advantage from multicellular organism. So beginning of the match goes purely situational, and it is the most important part, because more biomass you will accumulate in a first few thousand turns, the easier it will be later. Some sort of formation may be useful for fighters, but only untill they aill meet a missile. active dodging will also quickly break any formation. and fire concentration can be done without any formation. for economical units formation is not practically possible because they have to constantly move to avoid enemies and collect food. they cant share food so they have to check that they are taking only food that is not closer to other hungry ally gatherer. and as number of gatherers and food around grows, the complexity of such calculation grows exponentially. Calculation complexity is the main reason why gatherers shouls be as big as possible. Another is that small gatherers, with a size less than 10, cant properly stand on top of the food if engine is controlled directly. In short gatherers make only problems to each other when they are close.